ResearchTasksMulti-TaskMulti-TaskFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedGPQAIBTPO41.7Avg@32 Accuracy16May 28, 2026LongBench ChatDPO w/ LongReward72.6Point-wise Rate10Feb 26, 2026SimulationRLDC70Success Rate5Jun 23, 2026Real-world (test)RLDC40Success Rate5Jun 23, 2026